GRID - Host Group - Done and Exited Jobs

This graph shows the absolute counts of the number of done and exited jobs that are scheduled to hosts in a host group and how they are changing with time. The vertical axis represents the absolute job count while the time-based axis is selectable as is the case with other RTM graphs.

In the following example, the weekly chart has a modest rate of completed jobs followed by a spike in the number of exited jobs. In this example the unusual spike is the result of a job array that consists of over 2,000 job elements was ended by the LSF administrator. A day later, the jobs array was resubmitted and there is a steady increase in the rate of completed jobs. The reason for the decrease and apparent plateaus in the number of done and exited jobs has to do with the CLEAN_PERIOD setting in the lsb.params configuration file that governs how long mbatchd continues to retain information about recently completed or exited jobs.

Figure 1. GRID - Host Group - Done and Exited Jobs